PREDICTORS OF ELDERLY SUBJECTIVE WELL-BEING IN HUMAN ECOLOGICAL
PERSPECTIVE
By
Nan Zhang

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Human Development and Family Studies
2010

 
 

ABSTRACT
PREDICTORS OF ELDERLY SUBJECTIVE WELL-BEING IN HUMAN ECOLOGICAL
PERSPECTIVE
BY
NAN ZHANG
In recent decades, the well-being of elderly has drawn considerable attention especially
on subjective well-being. Many social topics have been studied from different environmental
levels in ecological perspectives, but elders’ subjective well-being has not yet been discussed
with multilevel environmental factors. The primary objective of this study was to investigate the
predictors for subjective well-being of American elderly on three levels: individual
characteristics, family attributes, and neighborhood social relationships. Data from General
Social Survey 2008 cross-sectional was used for analysis. Ordinal logistic regression was used to
examine the effects of selected individual, family and neighborhood variables on subjective wellbeing of elderly.
The results indicated that married elderly in the sample were more likely to have higher
subjective well-being than unmarried group. The analyses also demonstrated that family
financial situation and frequencies of church participation were strong predictors of elderly
subjective well-being controlling individual, family and neighborhood related factors. The
findings in the study will contribute to gain a better understanding of influential factors of elderly
subjective well-being among all levels variables, and will help social service professionals and
policy makers in their effort to improve the quality of life of the elderly. 

 
 

ACKNOWLEGEMENTS
I would like to express my sincere gratitude for the support of the members of my
committee, Drs. Robert J. Griffore, Lillian A. Phenice, and Lawrence B. Schiamberg. I
am especially grateful to Dr. Griffore for his guidance, as well as the extra time and
patience on me. This thesis could not have been done without Dr. Griffore’s time
devoting and crucial feedbacks.
I would like to thank Dr. Lillian A. Phenice and Dr. Lawrence B. Schiamberg for
being my committee members and providing their support and suggestions on this paper.
And I would like to extend my appreciation to Yun-Jia Lo who is my statistic consultant
in the CStat. She has provided helpful data analysis and interpretation advices.
I must acknowledge my younger brother who is always there with me. He is my
motivation to pursue the academic life, he is my motivation to challenge myself, and in
particular, to complete the master’s thesis.

(Major Professor: Robert J. Griffore
Committee: Lillian A. Phenice
Committee: Lawrence B. Schiamberg)

iii 
 

TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES .......................................................................................................... vii
CHAPTER 1
INTRODUCTION
Statement of the Problem .........................................................................................1
Purpose of the Study ................................................................................................2
Significance of the Study .........................................................................................3
Conceptualization of Subjective Well-being ..........................................................4
Measurement of Subjective Well-being...................................................................5
Research Questions and Hypotheses .......................................................................7
Human Ecology Theoretical Approach ...................................................................8 
CHAPTER 2
LITERATURE REVIEW
Individual and Demographic Factors ....................................................................12
Family Factors ......................................................................................................14
Neighborhood Indicators .......................................................................................16
CHAPTER 3
METHODOLOGY
Data Source ............................................................................................................19
Measures ................................................................................................................20
Data Analysis .........................................................................................................22
CHAPTER 4
RESULTS
Descriptive Analyses .............................................................................................23
Test of Main Hypotheses .......................................................................................27
Model Fitting Information .....................................................................................31
CHAPTER 5
DISCUSSION
Key Findings ..........................................................................................................33
Limitations .............................................................................................................36
Implications for Children’ Subjective Well-being .................................................38
Implications for Future Research and Policies ......................................................40
Conclusions ............................................................................................................42
APPENDIX A ....................................................................................................................44
APPENDIX B ....................................................................................................................46

iv 
 

APPENDIX C ....................................................................................................................48
REFERENCES ..................................................................................................................50 

v 
 

LIST OF TABLES
Table 1. Characteristic of Respondents for Categorical Variables ....................................25
Table 2. Results from Ordinal Regression Models Predicting Elderly Subjective
Well-Being ...........................................................................................................30
 

Table 3. Test of Parallel Lines ...........................................................................................32
Table 4. Goodness of Fit ....................................................................................................32
Table A 1. Questions Employed in the Analyses ..............................................................45
Table B 1. Missing Values for Variables Used in Data Analysis ......................................47
Table C 1. Available Link Functions in the Ordinal Regression Procedure in SPSS ........49

vi 
 

LIST OF FIGURES
Figure 1. Factors examined on individual, family and neighborhood levels .....................11
Figure 2. Distribution of General Happiness .....................................................................28

 
 

vii 
 

 

CHAPTER 1

INTRODUCTION

Worldwide aging as a social and a public issue has prompted much research on
the health of the elderly. The measurement of elderly health is not limited to objective
measures such as physician’s report. Subjective measures have also been taken into
account, for example, self-rated health, life satisfaction and happiness. Subjective wellbeing is one of the indicators of elderly overall health, and as an indicator of health it
contributes to the research of quality of life of the elderly.
Statement of the Problem
Population aging is a serious social issue across the world. The percentage of the
elderly has risen sharply in the U.S. The low mortality of older adults combined with the
baby boomers cohorts implies a continuous increase of the aging population in the 21st
century (Robine, Jagger, Mathers, Crimmins & Suzman, 2003). According to the U.S.
Bureau of Census, the population of those 65 years and over in the United States was
35.6 million in 2002, which comprises 12.3% of the total U.S. population. In the 20062008 the American Community Survey, it reported that people aged 65 and over were
37.9 million, which is 12.6% of the total population. It is estimated that in the year 2030,
the population of 65 and older will double to 71.5 million persons or 20% of the total
U.S. population (Young, 2004).
Given that this segment of the population is large and growing, it is important to
know if elderly perceive that they are satisfied with their quality of life, and whether they
1 
 

 

are happy. A question of increasing interest relative to the demography of aging is
whether or not Americans are living higher quality as well as longer lives (Yang, 2007).
Previous research showed that elders’ happiness remains stable throughout the years
(e.g., Easterlin, 2001).
There is substantial support that subjective well-being encompasses many
variables. According to the literature, the most influential variables include health status,
having good social relationships, access to local amenities, living in a safe, friendly
neighborhood, having access to transportation, having financial capacity to meet needs,
and having the ability to participate in the community (Cantarero, Potter & Leach, 2007).
Factors like health, income, and social relationships have been identified as predictors of
elder’s subjective well-being.
However, most of the predictors studied in the literature are at the individual
level. Little attention has been given to the association of elder’s subjective well-being to
family and other contextual factors. It is important to consider the family and other
environments in which people primarily live. Are there family variables and other
ecological variable that make influence elders’ subjective well-being?
Purpose of the Study
Among many factors which have been studied previously, there are no studies
that identified the different ecological levels found in understanding the subjective wellbeing of the elderly. . From my perspective, it is important to use an ecological system to
group these important factors from the different levels of the environment.
2 
 

 

The purpose of this study is to explore the influential factors from an ecological
perspective, especially the microsystem variables, to see how the interaction with one’s
family and other social group effect on individual’s subjective well-being and how they
vary from the individual traits themselves. This study is designed to examine whether
selected individual, family, or neighborhood variables can explain elders’ subjective
well-being and to examine the extent to which environmental variables explain elders’
subjective well-being compared to the influence of individual characteristics.
The purpose of studying elders’ subjective well-being under human ecological
perspective is to identify the conditions influencing subjective well-being. Family is the
primary environment in which individuals live. As the American family has changed in
complex way in recent years, it is important to study the current family system and its
influence on elder people’s subjective well-being, as well as the neighborhood context of
families.
Significance of the Study
The study of subjective well-being in the older population can benefit an
individual’s overall health. Being aware of the elderly’s perceived well-being is as
important as an objective status of health. Both physical health and subjective well-being
are important components of life quality and successful aging. Rowe and Kahn (1998)
defined three criteria for successful aging: low levels of disease and disability, high
physical and cognitive functioning, and active engagement with life. However, a large
number of older adults report high levels of well-being despite physical, cognitive and

3 
 

 

social deficits (George, 2006).Therefore, research in subjective well-being can influence
the literature as well as the scholarship of what constitutes successful aging.
This study provides a human ecological framework for better understanding of
elderly subjective well-being. Specifically, it addresses the individual, family, and
neighborhood variables, and points out the different levels of environmental factors.
Important predictors among related factors beyond just finding the association between
variables are identified for future research.
The research of elder’s subjective well-being is useful for the evaluation of
policies in many domains, including health care, public health, social service, families
and environment. In democratic societies such as the U.S., the indicators of subjective
well-being provide an important source of information to leaders about the well-being
and concerns of the older citizens (Diener, 2006).
Conceptualization of Subjective Well-being
Subjective well-being incorporates life satisfaction, happiness, affective
experiences, quality of life which tend to be used interchangeably in the research of
subjective well-being. The literature suggests that the measures of these concepts are
highly correlated among each other and with relevant personal and social characteristics
(George, 2006). In this particular study, subjective well-being will represent the meaning
of subjective perception of life.
Subjective well-being refers to the various types of evaluations people make
regarding their lives. Usually, it includes cognitive and emotional evaluations. Following
4 
 

 

Diener’s (1999) conceptualization, it consists of three major components: life satisfaction
as a cognitive aspect; the frequent experience of positive affect (PA); and the low
frequency of negative affect (NA). Both PA and NA have been considered as affective
aspects. The conceptualization of subjective well-being as composed of three factors,
namely, PA, NA, and life satisfaction, has received consistent empirical support (Lucas,
Diener & Suh, 1996; Arthaud-Day, Rode, Mooney, & Near, 2005).
Some researchers claim that subjective well-being need not be constructed so
narrowly (Russell, Burton, Rushing, Ritter & Rakocy, 1993). Later on, Diener (2006) has
added specific domain satisfactions: work satisfaction, marital satisfaction, or satisfaction
with other relationships.
However, there are alternative conceptualizations, such as happiness, morale and
satisfaction. In this study, subjective well-being is applied to represent the individual’s
subjective perception of life in terms of general happiness. It is generally agreed that
happiness is a subjective, positive, and inner psychological state of mind (Veenhoven,
2010). In this sense subjective well-being is a concept for the individual to assess whether
one’s life is good.
Measurement of Subjective Well-being
Measures of subjective well-being are obtained through self-reports. People are
asked to evaluate their lives as a whole or some aspect of it (Van Hoorn, 2007). In the
earlier research on subjective well-being, researchers studying the facets of happiness
usually relied on only a single self-report item. For example, the item in General Social
5 
 

 

Survey on overall happiness is: “Taken all together, how would you say things are these
days-would you say that you are very happy, pretty happy, or not too happy?”
Recent measures of subjective well-being contain multiple items. The Satisfaction
with Life Scale (SWLS; Diener, Emmons, Larsen & Griffin, 1985) and the Positive and
Negative Affect Schedule (PANAS; Watson, Clark & Tellegan, 1988) are the much used
ones. There are several statements that the respondents are asked to indicate the extent to
which they feel this way. SWLS asks participants to scale their agreement on 1-7 points
on five statements, such as “In most ways my life is close to my ideal”. PANAS measures
positive affect and negative affect each with 10 items. There are other scales that have
been evaluated which measure subjective well-being accurately (Kashdan, 2004): such as
the 4-item Subjective Happiness Scale (Lyubomirksy & Lepper, 1999), the 18-item WellBeing Scale (Tellegen, 1982). The literature suggests that the decision to use a single
item scale or a multi-item scale for measuring subjective well-being depends on the
researchers’ needs and the research feasibility.
For the measurement of subjective well-being in using the data from General
Social Survey (2008), researchers used a single-item self-report variable. Despite the
simplicity of subjective well-being measure, it has adequate validity and reliability
(Veenhoven, 1996). Besides, multi-items including domain satisfactions have been
applied to measure subjective well-being (Ellison, 1991). Specifically, there are items
related to satisfaction with community life, satisfaction with nonworking activities,
satisfaction with family life, satisfaction with friendship and satisfaction with health.

6 
 

 

However, the domain satisfaction variables in GSS are not available in recent
updated dataset; the most recent year containing domain satisfaction variables is 1994
which might not explain the present situations. Even from the dataset of 1994, the data of
domain satisfactions have over 50% missing data for the old population who are 60 years
and over. Therefore, in the present study, a single-item found in 2008 dataset will be used
to measure subjective well-being due to its availability and validity.
Research Questions and Hypotheses
Several variables are hypothesized as predictors of elders’ subjective well-being.
Variables will be clustered into three groups: individual level, family level and
neighborhood level. Individual level variables are social class, work status and education
attainment. Health has been identified as a predictor of subjective well-being commonly.
However, due to the lack of availability of health in all samples (sample design detail can
be seen in descriptive analyses section) in the General Social Survey, the variable health
was not included in the study. Marital status, family size and family economics will
describe the family level characteristics. Neighborhood social relationships will include
the frequency of social contact and church attendance.
The research questions here are as follows:
(1) How do three different levels of factors predict elders’ subjective well-being?
(2) Would the environmental microsystem variables (family and neighborhood)
be more influential than individual characteristics in predicting elders’ subjective wellbeing?
7 
 

 

As with the emphasis of individual traits and environmental factors in human
ecological perspective, the study hypothesizes that both family and neighborhood
variables are associated with elders’ subjective well-being as well as the individual
factors. Family and neighborhood factors may contribute to elders’ subjective well-being
over the individual factors.
The hypotheses posed in this study are that each of the variables examined does
not predict elderly subjective well-being. The following are these variables: social class,
education, work status, marital status, family size, family economics, social contact and
church attendance.
Human Ecology Theoretical Approach
Human ecology theory is unique in its focus on humans as both biological
organisms and social beings in interaction with their environment (Bubolz & Sontag,
1993). The quality of human life and the quality of the environment are seen as
interdependent, and a person’s behavior is a function of interaction between individual
traits and the environment. Here, elder’s subjective well-being will be studied for part of
the quality of life, in terms of person’s feeling of happiness.
Humans and their interaction with an environment constitute an ecosystem.
Bronfenbrenner (1993) identifies four major environmental levels in the ecosystem: the
microsystem, the mesosystem, the exosystem, and the macrosystem. It ranges from very
intimate contact between individuals to the broader social or physical context, like
government or weather.
8 
 

 

Microsystem is the environment with which individuals have intimate physical
and social interactions. Like the family or workplace are the places old adults having
their most activities and interactions. These intimate and interpersonal interactions in the
immediate environment are also referred as “proximal process”. Mesosystem consists of
the interactions among several microsystems, for example the interactions between
family and workplace. A fundamental ecological assumption is that what happens in a
person’s microsystem is interrelated to and interacts with other microsystems (Muss,
1996). The exosystem is a larger setting than the mesosystem and has direct or indirect
influence on individual, even without the individual’s participation. For example, the
welfare system and city facility are having effects on a disabled person’s daily life.
Macrosystem is an even larger context than exosystem, like a country’s general culture
and economic context. The great depression had adverse developmental outcomes for
individuals who born at that time. Therefore, the four levels of environmental contexts
construct the whole ecosystems in which an individual engages in physical and social
activities. For the present study of elders’ subjective well-being, microsystems of elderly
will be the focal scope.
In Bronfenbrenner’s view, proximal process is essential to the process of
development, without adequate proximal process, optimal developmental outcomes will
not occur (Griffore & Phenice, 2001). A microsystem consists of a cluster of activities,
roles and interpersonal relations experienced by the individual in a given face-to-face
setting with particular environment (Bronfenbrenner, 1993). Among microsystems,
family is considered as the primary environment. An individual’s well-being cannot be
considered apart from the whole family’s circumstances. A family carries out physical9 
 

 

biological sustenance, economic maintenance, and psychosocial and nurturance functions
for its members and itself (Bubolz & Sontag, 1993). But the family does not support its
members alone. Families interact with other microsystems through energy, resources and
information transformation. Neighborhood and church would be the other typical
microsystems for elders. Individuals are interacting in those settings, and these
microsystems as individual’s proximal environmental contexts have effects on
individual’s behaviors and well-being.
In the consideration of the effects of elder’s subjective well-being, only primary
environmental context is not enough. Characteristics of the individual also play an
important role in the process of interaction with environments. According to human
ecology theory, the organism does not receive environmental influence passively.
Personal characteristics can facilitate or impede the interaction with environment by
shaping and influencing proximal processes, persons who are active or very attractive
tend to elicit certain types of responses from social environment (Griffore & Phenice,
2001). The organism and environment affect each other reciprocally.
The human ecology perspective shares the symbolic interaction approach of
subjective interpretation, and perception is a significant factor in viewing the interaction
process. It is believed that environment should not be viewed only as the objective
external conditions in which individual lives. Rather it must be understood from the
individual perspective with respect to the meaning they have created based on their
needs, values and goals. People respond selectively to environmental stimuli and then
symbolically interpreted through personal and cultural meanings (Bubolz & Sontag,
10 
 

 

1993). Specif
fically, the objective nice and safe ne
e
eighborhood might attra more outs
d
act
side
ac
ctivities, but people who perceive vi
t
o
iolent or thre from the neighborhoo may
eat
od
su
ubjectively hold back their interactio
h
ons.
Multid
dimensional and recipro causality characteriz relationsh and
l
ocal
y
ze
hips
ex
xplanation in an ecosyst (Griffore & Phenice, 1988). So a multidimen
n
tem
e
nsional mod is
del
needed to und
derstand and illustrate in
d
ndividual’s subjective we
ell-being in t ecosyste
the
em.
Under a huma ecologica perspectiv subjective well-being of elderly w be studie in
U
an
al
ve,
e
g
will
ed
multidimensional levels: individual c
m
characteristic family fac
cs,
ctors, and ne
eighborhood
d
in
ndicators. In other word individual characteris
n
ds,
stics, family attributes, an neighbor
nd
rhood
variables will be examine as the stru
l
ed
ucture which are impacti on the in
h
ing
ndividual’s
ou
utcomes-eld
ders’ subjecti well-bein All facto on these t
ive
ng.
ors
three levels w presen
were
nted
in Figure 1.
n
Figure 1
e
Factor examined on individu family an neighborh
rs
d
ual,
nd
hood levels

Individual
• Ed
ducation
• So
ocial class
• Work status
W

Family
y

Neighb
borhood

• Marital stat
tus
• Family size
e
• Family econ
nomics

• Social co
ontact
• Church a
attendance

11 
 

 

CHAPTER 2

LITERATURE REVIEW

Many investigations in the literature have been designed to discover the
determinants of subjective well-being. Diener and his colleagues have examined on the
determinants of subjective well-being for years. Wealth, political and civil rights, social
comparisons, equality, culture traits have been researched in predicting subjective wellbeing of nations (Diener et al., 1995).
Larson (1978) has reviewed the previous thirty years of research on subjective
well-being of older Americans. In his work, he reported the related factors to subjective
well-being in the literature: Health, physical disability, socioeconomic status, age, sex,
race, employment, marital status, residence and social interaction. Most recently, George
(2010) reviewed literature since 2005 for subjective well-being in the later life. She
indicated that over fifty variables have been tested as determinants of subjective wellbeing. The major factors have been identified are: Health, social integration, social
relationships and social support, and psychosocial resource.
Among all those indicators identified in the literature, this study will review the
important factors in three dimensions under human ecological framework: individual
characteristics, family microsystem characteristics, and neighborhood factors.
Individual and Demographic Factors

12 
 

 

Most individuals experience some decline of health condition as they enter old
age. Does health decline means less happiness in life? Different measures of health have
been studied for subjective well-being. Most studies use self-assessments of health,
typically asking, “In general, would you say your health is: very good, good, poor, or
very poor?” An analysis from the Berlin Aging study (BASE: Baltes & Mayer, 1999)
shows that self-rated health proved to be the strongest predictor of subjective well-being
(Smith et al., 2002). Compared to the chronic illness and functional health, subjective
health contributes greater to individual differences in life satisfaction and aging
satisfaction. Disability of health condition are more incline to constraint physical
activities and social activities, while subjective health influence individual’s perception
of overall well-being. On the other hand, happier people and those who are more satisfied
with their lives reported better health (Siahpush, Spittal & Singh 2008). Self-rated health
and subjective well-being are highly correlated.
Research has shown that educational attainment is significantly correlated with
subjective well-being in adulthood, but formal education has only a small influence on
older adults’ subjective well-being (Witter, Okun, Stock & Haring, 1984). Layard (2005)
also mentioned that education has a small direct effect on happiness, and education
increases happiness by raising income. Individual unemployment has a large negative
effect on subjective well-being (Clark & Oswald, 1994). No doubt employment status is
related with income, and unemployment brings income decrease. Pinquart and Sorensen
(2000) synthesized findings from 286 empirical studies They reported that socioeconomic
status explains 2.2% to 3.2% of the variance of subjective well-being, and only 3.2% to
4.4% of the variance of subjective well-being is explained by income in old age.
13 
 

 

Several studies have found a positive relationship between age and subjective
well-being (Campbell & Sawer, 1976; Fernandez & Kulik, 1981; Hong &
Giannakopoulos, 1994). Subjective well-being (especially life satisfaction) remains
stable throughout adulthood and does not show large negative age differences even at
very old ages (e.g., Diener & Suh, 1997; Smith et al., 1999). Old people are more
satisfied with their lives since most of them have achieved their expectations in later
years. Moreover, older adults are more likely to compare themselves with people whose
levels of well-being are considered to be lower than their own. This is called downward
social comparisons (Gana, Alaphilippe & Bailly, 2004).
Some research has found no gender difference in subjective well-being. However,
Inglehart (2002) found a gender difference in subjective well-being. Women whose age
was under 45 tended to be happier than men. In contrast, older women were less happy
especially in rich societies (Inglehart, 2002). Most investigators have found that
race/ethnicity is not significantly related to subjective well-being (e.g., Campbell &
Sawer, 1976). Even when significant bivariate racial/ethnic differences are observed, they
usually disappear when socioeconomic status is taken into account (e.g., Brown, 1988;
Krause, 1993).
Family factors
Family is a fundamental unit of society. It serves both family members and
society; it provides economic support and protections to vulnerable members (Ooms,
1996). Family structure and family economic are commonly related to family members’
well-being and attainment. Research has emphasized the relationship between family
14 
 

 

factors and life satisfaction in childhood, adolescence, and adulthood. For example,
research has examined the relationship between marital status and subjective well-being.
For general population of older people, studies show married people to have higher
average subjective well-being score (Larson, 1978). Along with changes in health,
changes in marital status were the strongest predictors of changes in subjective wellbeing (Mroczek & Spiro, 2005). Other cross culture research has shown that being
married was strongly and significantly related to higher subjective well-being (Diener,
Gohm, Suh & Oishi, 2000). The sample was from 42 nations, and median sample size
was 1,027 per nation. They also concluded that the relations between marital status and
subjective well-being are similar across the nations. It has also been found that when age
was controlled, marital status was a less powerful predictor of subjective well-being
among people age 65 and older than among those between 40 and 64 years old (George,
Okun & Landerman, 1985).
Some conventional knowledge holds that money can buy happiness. In reviewing
30 years of research on subjective well-being of older Americans, Larson (1978) showed
the established relationship between socioeconomic status and subjective well-being. And
the correlations are consistent in showing mostly modest correlations between income
and subjective well-being (Diener & Diener, 2002). Older persons of lower
socioeconomic status tend to have lower subjective well-being. Therefore, family
economics is considered to be a factor of prime importance to the welfare of the elderly
(Streib & Beck, 1980), which directly influence on their satisfaction with life.

15 
 

 

Recent theory also demonstrates other propositions regarding the relationship
between income and subjective well-being (e.g., Diener et al., 1999, Cummins, 2002).
Cummins proposes that money buys happiness to the extent that external resources
permit optimal functioning of the subjective well-being. Level of income can decrease or
increase with an individual’s availability of living resources. Poor people with poor
nutrition and limited medical care increase their possibility of illness and disability. Poor
people living in a more violent neighborhood are likely to experience bullying, fear and
insecurity. They are likely to have more limited life choices. All these conditions may
lead to low life satisfaction with negative influence on subjective well-being. Therefore,
income is a strong predictor of subjective well-being, especially for the poor population.
However, income becomes less related to subjective well-being when people have
sufficient resources to support basic needs. Their needs go beyond the resources which
money can buy, associated with Maslow’s hierarchy of needs. A well known Easterlin
(2001) wealth-happiness paradox study concluded the same result that additional
increases in wealth do not result in higher levels of happiness.
Neighborhood indicators
Neighborhood is a very important microsystem in which people live. The
neighborhood environment influences social interaction directly, which potentially
effects on people’s satisfaction of daily life. Neighborhoods contain two different types
of environments: physical environments and social environments. Pearlin and Skaff
(1996) found that as people aged they may become increasingly wary of the quality of
their community and neighborhood environments for their safety.
16 
 

 

Therefore, living in a safe, friendly neighborhood is very important, as reported
by the elderly themselves (Cantarero, Potter & Leach, 2007).
Individuals have enduring social interactions with neighbors. Neighbor’s life
condition is also another potentially important factor. One way that neighbors can have
effects on an individual’s subjective well-being is through social comparison. The idea is
that individuals will be happier if their income is higher than people lived nearby, but will
be less happy if they are below the reference standard (Diener et al., 1995). For the
standard of comparison, there are upward social comparisons and downward social
comparisons. For the upward comparison, people who are likely to compare themselves
to someone more advantaged are inclined to view lives negatively. In the contrast, people
who view lives positively are likely to choose someone less advantaged to compare.
Research indicates that older adults are more likely to use downward social comparisons
than are young and middle-age adults (George, 2006).
Along with the decreased work or retirement, social interactions decline over old
age, church may become an important place for people again. After children grow up and
leave home, the house became an empty nest, which means parents have more time for
activities beyond families. Religious participation, specifically attending religious
services, is a strong predictor of subjective well-being for older adults (Ellison et al.,
2001; Warr et al., 2004). Also Krause (2003) has reported that finding meaning through
religion is positively related to subjective well-being. Religion and the church community
can be a source of social activities and support and also possibly a comfort during

17 
 

 

stressful times. Evidence has been found in support of the stress-buffering effects of
religious beliefs enhance life quality (Kirby et al., 2004).
Actively religious people have reported markedly greater happiness and life
satisfaction than irreligious people (Ciarrochi & Deneke, 2005). Religion in people’s life
has been studied in spiritual beliefs and church attendance. Seligman (1988) has argued
that a loss of meaning underlies today’s high depression rate, and that finding meaning
requires an attachment to something larger than the lonely self.
A study of happiness in the United States showed 43% of weekly or more church
attenders and 26% of seldom or never church attenders reporting themselves “very
happy” (Pew, 2006). In National Opinion Research Center (2006) surveys, 40% of people
who feel extremely close to God feel very happy, compared to 24% of those who do not
feel close to God. There are no marked differences in happiness by religion. About one in
three Protestants, Catholics, and Jews have reported being very happy (Myers, 2008).

18 
 

 

CHAPTER 3

METHODOLOGY

This is a cross-sectional study of elders’ subjective well-being from the human
ecological perspective. Individual characteristics and intimate social context will be
examined. Family is the primary microsystem in which elders interact physically and
socially, other microsystems like church and neighborhood also play important roles in
elderly well-being.
Data Source
The data for this research is from the General Social Survey (GSS). The GSS is a
personal interview survey of U.S. households conducted annually by the National
Opinion Research Center. It is repeated cross-sectional surveys from 1972 to 2008, and
provides the best national-representative source of data on happiness and is part of the
World Database on Happiness (Yang, 2008). There is a wide range of questions including
demographic, behavioral, and attitudinal questions, with a sample size that ranges
between 1,500 and 3,000 respondents in each wave. The response rate over years ranges
from 73% to 79% (Louis & Zhao, 2002).
The present study is based on the most recent available subset of GSS in 2008.
The total sample size of this subset is 2,023, and consists of 529 cases of people aged
over 60 years, including 232 males and 297 females. Most respondents in the sample are
white.

19 
 

 

Measures
Subjective well-being is assessed in the GSS as a single question about general
happiness: “Taken all together, how you would say things are these days-would you say
that you are very happy, pretty happy, or not too happy?” The responses are coded as 1,
2, and 3 respectively. In this research the responses were recoded so that 3 represents
very happy, 2 represents pretty happy, and 1 represents not too happy. I would have been
preferable to use a multiple item scale to measure subjective well-being, but because
secondary data analysis is utilized in this study, it is not possible to reconstruct a
subjective well-being scale out of the GSS. However, this is one of the most commonly
used measures for subjective well-being or happiness in the United States. Moreover, it
has been found that single item measures of happiness are both valid and reliable
(Crooker & Near, 1998).
There are three levels of independent variables to predict elders’ subjective wellbeing: individual, family, neighborhood.
The individual level characteristics of elderly included social class, work status,
and degree. The measure of social class asks: Which would you say you belong in: the
lower class, the working class, the middle class, or the upper class? The responses were
coded as: (1) lower class, (2) the working class, (3) the middle class, and (4) the upper
class. Work status is a six category variable: full-time worker, part-time worker,
unemployed, retired, keeping house, and other. It was recoded into four groups: (1) not
work or other, (2) retired, (3) working part time, and (4) working full time. Education
was measured by degree. Responses included (0) less than high school, (1) high school,
20 
 

 

(2) associate/junior college, (3) bachelor’s, (4) graduate. The responses were recoded
from 1-5 (eliminating zero code) to be consistent with other variables.
Family level variables were marital status, family economics, and family size.
Marital status was asked by “Are you currently – married, widowed, divorced, separated,
or have you never been married?” It was recoded into (1) married and (2) unmarried.
Family economics was assessed in terms of family financial situation, which was
measured by asking “so far as you and your family are concerned, would you say you are
(1) pretty well satisfied with your present financial situation, (2) more or less satisfied, or
(3) not satisfied at all.” The categories of responses of the two variables were reverse
coded as well. Family size was asked by how many family members in the household.
Neighborhood social environment was measured by social interaction with
neighbors and church attendance. Social interaction within neighbors was asked by “How
often you spend a social evening with someone who lives in your neighborhood”?
Responses categories consist of (1) almost every day, (2) once or twice a week, (3)
several times a month, (4) about once a month, (5) several times a year, (6) about once a
year, and (7) never, and also was recoded in ascending order.
The interaction with church was measured by church attendance. Church
attendance was asked by “how often do you take part in the activities and organizations
of a church or place of worship other than attending services?”, response categories
include (1) never, (2) less than once a year, (3) about once or twice a year, (4) several
times a year, (5) about once a month, (6) 2-3 times a month, (7) nearly every week,
(8)every week, (9) several times a week, (10) once a day, (11) several times a day.
21 
 

 

Data Analysis
SPSS 17was used to analyze the data. First, descriptive statistics were calculated,
including frequencies, percentages, means and standard deviations, as appropriate, to give
a general description of all variables used in this study. And the number of missing data
for every variable was also be given; missing data less than10% was considered as
missing at random (Shrive et al., 2006).
Second, chi-square analysis was used to test the relationships between each
categorical independent variables and dependent variable. While one way ANOVA f test
was used to test the association between continuous independent variables and the
response variable.
Third, ordinal regression was used for the main hypothesis testing. Ordinal
regression is appropriately used when the dependent variable is ordinal which is indicated
by ranked categories. Here, the dependent variable general happiness has three
categories: very happy, pretty happy, and not too happy. The response is considered to be
ordered from not too happy as the lowest level of happiness to very happy. Ordinal
regression in SPSS is an extension of the general linear model to ordinal categorical data.
There are three levels of predictors, and the purpose of this study is to test the interaction
between subjective well-being and three levels of social context. First, the demographic
and individual characteristic was put into the model. And then all predictor variables
from three ecological levels were put into the model together to see which variables
predicted the elder’s subjective well-being.

22 
 

 

CHAPER 4

RESULTS

This chapter describes the findings from this study. Two sections are included,
first descriptive analyses of all variables considered and the association between
dependent variable and all independent variables in the study are presented, and then the
results of the ordinal regression analysis are exhibited to predict the elderly subjective
well-being from the set of individual, family and neighborhood level factors.
Descriptive Analyses
The original sample (N=529) was obtained from the General Social Survey crosssectional 2008 on condition that respondents were aged 60 and over. However, the
sample had to be reduced due to the design of GSS. There were three random subsamples: A (N=171), B (N=160) and C (N=198). Sample C did not include the variable
social contact frequency. However, social contact frequency was an important variable in
the study, and the main hypotheses could not be tested without it. Therefore, sub-sample
C was eliminated from the study. Health was only asked for sub-samples A and C.
Therefore, health was excluded from the analyses. Since they were random sub-samples,
the two sub-samples (A and B, N=331) were used in this study. In addition, 10 cases
were discarded in the analyses because of having missing values in variables (See
appendix B). Finally, complete information was retained from 321 respondents.
Table 1 presents descriptive statistics related to categorical variables including
frequencies and percentages in each response group, also associations between general
23 
 

 

happiness and categorical independent variables are presented in this table. The majority
of respondents seemed to be pretty happy (54.0%) with life overall; however 33.5% were
very happy and 12.5% were not too happy. Demographic factors (i.e. age, gender and
race) have been controlled in the study. Among respondents from all ages, mostly were
from 60 to 79 age range and took 85.7% of the whole sample. The sample was 46.7%
male and 53.3% female with a mean age of 69.84. And 53.3% were married while 46.7%
were unmarried. With regard to ethnicity, the sample included 85.0% white, 11.5% black,
and 3.4% other ethnic groups. Most of the respondents were retired (56.4%).
Associations between categorical variables were carried out by Chi-square tests of
independence. Gender, race, degree, work status, social class, marital status, family
economics, social contact frequencies and church attendance were tested with general
happiness. Among all nine independent variables, the result showed that marital status,
family economics and church attendance were significantly associated with elders’
subjective well-being. As can be seen by the frequencies cross tabulated in Table 1, there
is a significant relationship between marital status and elderly subjective well-being, X2
(2, N = 321 ) = 20.353, p < .001. More respondents who chose pretty happy and very
happy were married than unmarried elderly. In contrast, the percentage of unmarried
respondents in the not too happy group was higher than that of married ones. When
looking at the family economics, different satisfaction with family financial situation
were not equally distributed in the population, X2 (4, N = 321) = 37.909, p < .001. Only
2.1% elderly where were satisfied with their family financial situation were in the not too
happy category, while 20.6% were very happy. For respondents who were not at all
satisfied, 1.8% more respondents were not too happy than who were very happy. For the
24 
 

 

religious participation, the percentage of respondents were differed by the frequencies of
church attendance, X2 (8, N = 321) = 18.247, p = .019. It is obvious that most people who
attend church activities every week were happy, only 3 persons were not too happy. For
the rest of the categories which are less than every week, most respondents were pretty
happy.
Table 1
Characteristic of respondents for categorical variables
Measures

General happiness
Not too
happy

N (%)
Age
60-69
70-79
80 and over
Gender
Male
Female

41(12.5)

Pretty
happy

Very
happy

175 (54.0) 105 (33.5)

Total

Pearson
Chi-square

321 (100)
57.869

23 (7.2)
13 (4.0)
5 (1.5)

102 (31.8) 51 (15.9)
48 (15.0) 38 (11.8)
25 (7.8)
16 (5.0)

176 (54.9)
99 (30.8)
46 (14.3)
4.574

18 (5.6)
23 (7.2)

74 (23.0) 58 (18.1)
101 (31.5) 47 (14.6)

150 (46.7)
171 (53.3)
5.458

Race
White
Black
Other

30 (9.3)
8 (2.5)
3 (0.9)

152 (47.4) 91 (28.3)
18 (5.6)
11 (3.4)
5 (1.6)
3 (0.9)

273 (85.0)
37 (11.5)
11 (3.4)

Degree
Less than high school
High school
Junior college
Bachelor
Graduate

14 (4.4)
14 (4.4)
3 (0.9)
6 (1.9)
4 (1.2)

23 (7.1)
98 (30.5)
10 (3.1)
26 (8.1)
18 (5.6)

54 (16.8)
171 (53.3)
18 (5.6)
47 (14.7)
31 (9.6)

12.542

25 
 

17 (5.3)
59 (18.4)
5 (1.6)
15 (4.7)
9 (2.8)

 

Table 1 Continued
General happiness
Measures
Not too
happy
Work status
Working full-time
Working part-time
Retired
Not work and other
Social class
Lower class
Working class
Middle class
Upper class
Marital status
Married
Unmarried
Family economics
Not at all satisfied
More or less satisfied
Satisfied
Social contact
Never
Once a year
Several times a year
Once a month
Several times a
month
Several times a week
Almost daily
Church attendance
Never
Once a year
Several times a year
At least once a month
Every week

Very
happy

8(2.5)
4(1.3)
18(5.6)
11(3.4)

36(11.2)
11(3.4)
99(30.9)
29(9.0)

13(4.1)
10(3.1)
64(19.9)
18(5.6)

57(17.8)
25(7.8)
181(56.4)
58(18.0)

5 (1.6)
18 (5.6)
17 (5.3)
1 (0.3)

20 (6.2)
64 (19.9)
87 (27.1)
4 (1.2)

5 (1.6)
32 (10.0)
62 (19.3)
6 (1.9)

13 (4.1)
28 (8.7)

85 (26.5)
90 (28.0)

73 (22.7)
32 (10.0)

171 (53.3)
150 (46.7)

Pearson
Chi-square

30 (9.4)
114 (35.5)
166 (51.7)
11 (3.4)

7.209

9.592

20.353***

37.909***
21 (6.5)
13 (4.0)
7 (2.1)

43 (13.4)
66 (20.6)
66 (20.6)

15 (4.7)
24 (7.5)
66 (20.6)

79 (24.6)
103 (32.1)
139 (43.3)

21 (6.5)
1 (0.3)
2 (0.6)
7 (2.1)
2 (0.6)
6 (1.9)
2 (0.6)

46 (14.3)
26 (8.1)
26 (8.1)
19 (6.0)
24 (7.5)
25 (7.8)
9 (2.8)

31 (9.7)
11 (3.4)
10 (3.1)
15 (4.7)
15 (4.7)
18 (5.6)
5 (1.6)

98 (30.5)
38 (11.8)
38 (11.8)
41 (12.8)
41 (12.8)
49 (15.3)
16 (5.0)

23 (7.2)
9 (2.8)
2 (0.6)
4 (1.2)
3 (0.9)

75 (23.4)
27 (8.4)
29 (9.0)
23 (7.2)
21 (6.5)

38 (11.8)
11 (3.4)
27 (8.5)
9 (2.8)
20 (6.3)

136 (42.4)
47 (14.6)
58 (18.1)
36 (11.2)
44 (13.7)

18.866

18.247*

26 
 

Total

Pretty
happy

 

Test of Main Hypotheses
The main hypotheses were tested using Ordinal Logistic Regression. Eight
variables were hypothesized as predictors of elders’ subjective well-being which was
measured by general happiness. When entering the blocks of predictors, there are two
options in SPSS 17. Factors are for categorical variables, while covariates are for
continuous variables. Therefore, two of them (age and family size) were entered into
covariates, while seven of them (subjective class, degree, work status, marital status,
family economics, social contact and church attendance) were entered into factors. Also,
the demographic variables have been taken in control, included age, gender and race.
Choosing the link function was a major decision before running the ordinal
regression. Probit link function was used in the model. It is crucial to choose link
functions in constructing the model to demonstrate the model appropriateness. The link
function is the function of the cumulative probabilities that results estimation of the
model. Five different link functions are available in the Ordinal Regression procedure in
SPSS: logit, probit, negative log-log, complementary log-log, and cauchit. For typical
application for each link function, see appendix C.
Looking at the distribution of outcome variable, general happiness, the middle
category pretty happy had the highest probability and there were no extreme values. So,
complementary log-log, negative log-log and cauchit link were not appropriate. The
commonly used logit link required equal categories of the dependent variable. Therefore,
probit link function was chosen to apply for the model since the general happiness was
closer to normal distribution, which is shown in Figure 2.
27 
 

 

Figure 2
Distribution of general happiness

Table 2 shows the results of the Ordinal Logistic Regression. Demographic
factors were controlled in the model. To test the effects of individual, family and
neighborhood factors, three demographic factors and three individual variables were
included in the model. Among these variables, there are three significant factors in
predicting subjective well-being of elderly.
Subjective well-being of elderly was significantly associated with three variables
(marital status, family economics, and church attendance). Of these three significant
predictors of elders’ general happiness, two (marital status and family economics) of
28 
 

 

them were on the family level while one (church attendance) was on the neighborhood
level.
Regression coefficients are shown in Table 2. Each category of every variable has
a coefficient, which means each category is compared to the last group respectively.
Form the observed significance level in the Table 2, you see that marital status, family
economics and church attendance are all predicting the elderly subjective well-being.
Compared to the unmarried group, Married (code 1) elderly are more likely to have
higher scores of subjective well-being (p = .001). Family economics and church
attendance both have negative coefficients. Family economics adequacy, which has three
categories (i.e., not at all satisfied, more or less satisfied, and satisfied), the first two
groups were compared to the last group. Elderly who are not at all satisfied with their
family financial situation are less likely to have higher scores on subjective well-being (p
< .001). This applies also to the elderly more of less satisfied with family economics (p <
.001). Church attendance is the other significant predictor positively related with
subjective well-being of elderly. Among all categories (i.e., never, once a year, several
times a year, at least once a month and every week), elderly who attend church activities
several times a year seems not related to subjective well-being (p = .834). While elderly
who attend church never (p = .047), once a year (p = .003) and at least once a month (p =
.027) tend to have lower subjective well-being compared to who attend weekly. Age
does not appear to be related to elderly subjective well-being.
In addition, the pseudo R square measured the success of the model in explaining
the variances in the data. The interpretation of pseudo R square in the ordinal regression
29 
 

 

model was similar to that of the R square in linear regression. The pseudo R square
indicated that the proportion of variances in the outcome variable was accounted by the
explanatory variables; the larger the pseudo R square, the better the model fitting (Chen
& Hughes, 2004). Based on Nagelkerke Pseudo-R2, 29% of variance was explained by
the model with individual, family and neighborhood factors.

Table 2
Results from ordinal regression models predicting elderly subjective well-being
Predictor

Estimate

 

.901
.574

.380
.424

.953
.997

.187
.491
.351
-.075

.296
.255
.373
.290

.527
.054
.348
.797

-.149
.344
.096

.245
.301
.193

.543
.253
.618

-.200
-.521
-.303

30 

.010
.139

-.022
-.002

Individual factors
[Degree=Less than high school]
[Degree=High school]
[Degree=Junior college]
[Degree=Bachelor]
[Degree=Graduate ]
[Work status=full-time]
[Work status=part-time]
[Work status=retired]
[Work status=not work and other]
[Class=Lower class]
[Class=Working class]
[Class=Middle class]
[Class=Upper class]

P Value

-.001
.078

Demographics
Age
[Sex=male]
[Sex=female]
[Race=white]
[Race=black]
[Race=other]

Std. Error

.459
.409
.393

.662
.203
.441

 

Table 2 Continued
Predictor

Estimate

Std. Error

P Value

.659
.190
Family level factors
[Marital status=married]
[Marital status=unmarried]
-1.026
.206
[Family economics=Not at all satisfied]
-.685
.166
[Family economics=More or less satisfied]
[Family economics=Satisfied]
.029
.138
Family size
Neighborhood level factors
[Social contact=Never]
-.308
.335
[Social contact=Once a year]
.058
.365
[Social contact=Several times a year]
-.322
.365
[Social contact=Once a month]
-.316
.361
[Social contact=Several times a month]
.068
.358
[Social contact=Several times a week]
-.086
.348
[Social contact=Almost daily]
[Church attendance= Never]
-.448
.225
[Church attendance= Once a year]
-.774
.262
[Church attendance= Several times a year]
-.055
.260
[Church attendance= At least once a month]
-.615
.277
[Church attendance= Every week]
R2
.290
Last group of categorical predictors is the reference group with no values.

.001*
.000*
.000*
.834

.359
.873
.377
.381
.850
.806
.047*
.003*
.834
.027*

R2 shown are Nagelkerke Pseudo-R2.
Model Fitting Information
When fitting an ordinal regression you assume that the independent variable
coefficients were equal across all levels of the outcome variable. That is, the slope for
class in “very happy” is the same as the slope for class in “pretty happy”. This is
confirmed in Table 3. The observed significance levels for both models were large
enough accept the hypotheses, which indicated the models were satisfied. More model
fitting information can be found from goodness of fit; the null hypothesis is that the
31 
 

 

distribution fits the data. The results can be seen from Table 4, the Pearson’s chi-square
indicated the overall model was fit.
Table 3
Test of parallel lines
-2 Log Likelihood

Chi-Square

Df

Sig.

.170

Null hypothesis

521.852

General

486.849

35.003

28

Chi-Square

df

Sig.

Pearson

616.717

606

.373

Deviance

519.080

606

.995

Table 4
Goodness of fit

Overall, the model fits well. Three out of eight variables were observed that they
significantly predicted elderly subjective well-being in the complete model, and they
were all family and neighborhood factors. Marital status and family economics are on
family level, and church attendance is on the neighborhood level.

32 
 

 

CHAPTER 5

DISCUSSIONS

In this study, the human ecological perspective was adopted as a foundation
for an analysis which focused on the elderly subjective well-being. Specifically, it
examined the family and neighborhood factors in addition to the individual
indicators which are found to be the most consistent factors in previous research.
Based on Bronfenbrenner’s view, proximal process is essential to the process of
development, the results of this study suggested that elderly family background
(financial situation) and the interaction with churches indicated the elderly
subjective well-being.
The discussions of the results are presented in three parts. The first part
examines key findings from the research, with discussions of factors important to
the changes and increases in elderly subjective well-being. The second part
discusses the limitations in the study. And the third part brought implications for
child development and future research.
Key Findings
This study describes the subjective well-being of elderly individuals and
identifies some critical factors to understand it. Overall, the findings support the
human ecological perspective for understanding factors that make an impact on the
subjective well-being of elderly at the individual, family or neighborhood levels.

33 
 

 

It is helpful to first look at the demographics. The characteristics of the
demographic data show that the sample is fairly representative. The final sample
size is 321, which is large enough for a national sample. Also the male and female
respondents were quite equally selected, so were the married and unmarried groups,
as shown in Table 1. There were 85% white elderly, 11.5% black elderly and 3.4%
from other ethnic groups. This was similar to the United States Census Bureau’s
race ratio report in 2009 (79.6% white, 12.9% black and 7.5% others). As a result,
the sample was able to represent the national elder population without oversampling any given race.
The mean score of the general happiness for elderly is 2.20 on a 3 point
scale, which is higher than the mean score (M=1.89) of the entire population. Less
work pressure has been identified as one of primary factors in higher subjective
well-being in later life. A role strain perspective states that retirement from the
demands of a job may reduce role strain, thereby enhancing subjective well-being
(Kim & Moen, 2001). In this study, only 17.8% of the respondents were working
full-time, and 56.4% were retired. We can see that the elderly tend to work less
than younger adults. But no significant relationship has been found between work
status and the elderly subjective well-being in this study.
This study has also found that marital status and family financial situation
on the family level were the most important predictors for elderly subjective wellbeing. Married elders tend to be happier than unmarried ones. The marital
relationship is one of the subgroups in the internal family system. This organizes
34 
 

 

the internal structure of the family system, which consists of individually
subsystems and their roles and relationships with each other (Griffore & Phenice,
2001). Regardless of marital status, the different internal family structure affects
the elderly family life experience and communication or information processing,
which directly influences their quality of life.
Family economics is another factor that has been discussed in previous
studies related to elderly subjective well-being. The association between family
economics and elderly subjective well-being has also been found in this study, and
it is one of the major predictors. Here respondents’ satisfaction of the family
financial situation was a subjective measurement in family economics. People who
were more satisfied with their family financial situation were more likely to have a
higher subjective well-being. From the human ecological perspective, money is a
type of energy flow (Odum, 1994). Higher income enhances subjective well-being
as it is beneficial to elderly fulfilling their basic needs (Diener & Diener, 2002).
Elderly use money as an exchange medium, which is the means to pertain resources
and satisfactions.
In the neighborhood level, this study found that religious participation is a
major and positive predictor of subjective well-being for older adults. Those who
going to church on a regular basis have better subjective well-being than those who
do not. This finding was consistent to the literature stating that religion tends to be
an especially powerful buffer against stress among elderly people (Boyden &
Aronui, 2001). Church is an important interactive agent in the neighborhood. It is
35 
 

 

also a great social interaction place besides family in the microsystem. It could as
well be seen as an external family system that influences the family atmosphere and
members in it. However, another variable of neighborhood level social contact was
not a significant predictor in this study. Social contact has been discussed as a
stronger predictor in previous studies.
In this study, age has not been found to be related to elderly subjective wellbeing. Some cross-sectional studies have reported that a positive relationship
between happiness and age. In the contract, longitudinal studies have indicated that
happiness showed a slight decrease after age 65 which is a peak. Moreover, age has
not been found to be associated with subjective well-being across different age
groups. People’s well-being is quite steady across the life span. It is not age alone
that significantly affects subjective well-being change, but rather significant life
events (Myers, 1992). Also mentioned in this study, significant life events and
related social activities are important to subjective well-being. Marital status is the
one of major factors in older adults’ quality of life. A healthy marriage is positively
related to elderly life satisfaction, as well as social activities in the community or
church play an important role in their subjective well-being.
Limitations
One of the limitations in this study was the use of a single item for
measurement of subjective well-being. Multi-item measurement has been
considered during the study. Specifically in General Social Survey, domain
satisfactions have been considered to be included in the measurement. Domain
36 
 

 

satisfaction refers to the particular satisfaction in different life domains, such as
satisfaction with city, satisfaction with hobby, satisfaction with family, satisfaction
with friends and satisfaction with health, presented in GSS data before 1994. Due
to the lack of domain satisfaction data in most recent dataset in GSS, it has not been
used in this study. This is an inevitable limitation for using secondary data
analyses.
On a consistent basis, family economics has significantly contributed to our
understanding of elderly subjective well-being. In this study, family economics was
measured by subjective satisfaction of the family financial situation. There were
also objective family income measures found to be positively related to the
subjective well-being (McLaughlin & Carter, 2007). Integrating the objective
family income with subjective family financial satisfaction will provide a better
understanding of family economics.
In this analysis, the pseudo R-square is not large. There might be other
variables not included that could have effect on elders’ subjective well-being. Due
to the constraint of using an existing dataset, not all related variables can be
included in this study. Health, which is a potential predictor of subjective wellbeing, was not included for reasons that have been discussed previously. What is
more, this study is focused on the microsystem factors, by using the ecological
perspective. Contextual factors associated with other levels of the ecosystem could
well be considered in future research. Social policies about health care could also
be an important factor, as well as the community facilities, social service
37 
 

 

organizations and the economic atmosphere. Therefore, out of the scope of this
study, more factors are encouraged to be included in future research.
Implications for Children’ Subjective Well-being
This study looked at the environmental factors of subjective well-being of
elderly. Another recent study addressed positive and negative life events as
determinants of subjective well-being in young, middle-aged and old adults from a
life span perspective (Gomez, Krings, Bangerter & Grob, 2009). The life span
perspective views that human development is a continuous growing process, the
behaviors in later life stages were still influenced by the past. This implies the
meaning of studying the subjective well-being on the early childhood development.
In research on the subjective well-being, for most studies, the age ranges from
teenagers to the elderly. Few studies have been focused on children. There are
some implications on studying children’ happiness generated from the study of
elderly subjective well-being under the human ecological framework.
First of all, children’ subjective well-being itself should be a concern of
future research. Research could address life satisfaction during the childhood years,
because children’ perceived happiness is as important as the objective living
environment for their quality of life. As previously shown the study focused on
elderly subjective well-being indicated that the significance of the family and
neighborhood factors. With regard to children’s happiness, it would be very
important not only to look at their personality or other biological characteristics.
Family background and community and neighborhood environment are also
38 
 

 

important. This study suggests that family and other proximate contexts might be
influential to the subjective well-being of children also, because family and
community are places children involve themselves and have close interaction with.
We have seen that the family financial situation is one of the most important
predictors of elderly subjective well-being. As described above, money is the
medium for information, materials and other resources flow. This is fundamental
for individual to attain satisfaction, both physically and mentally. Besides findings
from this study, another study focused on adolescence drew the same conclusion.
Discontent with the financial situation significantly reduced the quality of mental
health, leading to inappropriate patterns of behaviors, and endangers future wellbeing of adolescents (Pranjic, Brkovic & Beganlic, 2007). Therefore, family
financial situation might be also important to promote children’ subjective wellbeing. And the family structure, the internal relationship between family members
should also be included in the consideration of factors affecting children’
happiness.
As mentioned earlier, subjective well-being is quite stable across different
age groups, and age alone does not account for the changing of subjective wellbeing, whereas significant life events are likely to be more important. From the
findings in this study, it is clear that significant life events and related social
activities are important to elderly subjective well-being, this could as well be true
for children’ happiness. Obviously, significant events for children are quite

39 
 

 

different from older adults’ experience. For example, going to school, relocation, or
parents’ relationship change can be influential to their subjective well-being.
Moreover, grandparents caregiving could be an indicator of children’ wellbeing. Bronfenbrenner (1977) referred to the environmental systems closest to the
individual as the microsystems, which for most children, is very likely to be the
family and places such as school or their grandparents’ homes. In some families,
grandparents are taking responsibilities for their grandchildren. It is estimated that
39% of grandparents responsible for their co-resident grandchildren will be
responsible for their grandchildren for 5 years or more (U.S. Census Bureau, 2003).
So, subjective well-being of elderly is important to the happiness of children since
they are living together and influencing on each other, especially for those
grandparents caregiving families. Any changes in grandparents as a result of
caregiving are likely to result in changes in the environment of the children for
whom they are (Jooste, 2007).
Implications for Future Research and Policies
The findings in this study supported the utility of a human ecological
perspective for understanding factors influencing the subjective well-being of
elderly at the individual, family and neighborhood levels. Future research should
include these contextual factors affecting elderly subjective well-being. Future
research is needed on elders’ family environment to better understand the family
context affecting the elderly well-being. What are the family risk factors for elderly
subjective well-being? What are the exactly risk factors for unmarried elders? This
40 
 

 

study primarily studied on the microsystem factors. There are also mesosystem,
exosystem and macrosystems in the social ecosystems. Future research on elderly
subjective well-being including larger contextual factors is suggested to study this
topic under the full human ecological structure.
Experience sampling methodology (ESM; Diener, 2000) is an innovative
methodology for the assessment of subjective well-being. ESM involves the use of
palm computer technology to obtain a random sampling of a person’s mood and
cognitions over time, with this technology, researchers can obtain a record of the
ongoing subjective experience of the respondent without relying on the
retrospective reports of these subjective stats (Pavot, 2008). Thus, ESM is not
applied for this secondary cross-sectional data analysis, because it requires repeated
experiential reports throughout the day or week. A longitudinal design using ESM
is suggested for research in the future.
The findings in regard to family factors have provided implications for the
elder social services. An individual’s subjective well-being is related to his or her
relationship status and family economics. It is unlikely that happiness of the elderly
will be increased simply by expanding individual physical and mental activities.
Focusing on the family situation could also be important, like family financial
situation. As George (1992) pointed out, the integration of subjective measures
(financial satisfaction) and objective income measures may provide useful insight
for policy makers in approaching the issue of income adequacy for income
maintenance programs. And since subjective well-being is an important part of
41 
 

 

successful aging, the study of subjective well-being provides a criterion for
evaluations of policies in older population.
Conclusions
This study identified factors relating to elderly subjective well-being based
on a human ecological perspective. The scope of the study is limited to a certain
level of analysis at the microsystem (individual, family and neighborhood levels) in
order to concentrate on related factors mostly found in the previous research. The
results explained some significant family and neighborhood factors in addition to
the individual characteristics. I hope this study would bring attention to the
analyses of the characteristics of important factors in the literature. Much research
has been done primarily with the focus on finding the association between
subjective well-being and individual psychological characteristics. The major
purpose of this study highlights the importance of the human ecological perspective
to look at the elderly subjective well-being.

42 
 

APPENDIX

43 
 

APPENDIX A

44 
 

Table A 1
Questions Employed in the Analyses (Source: General Social Survey 2008)
Age. Respondent’s age. It was calculated by subtracting the year of birth from the survey
year. (AGE)
Church attendance. How often do you take part in the activities and organizations of a
church or place of worship other than attending services? (RELACTIV)
Class. If you were asked to use one of four names for your social class, which would you
say you belong in: the lower class, the working class, the middle class, or the upper class?
(CLASS)
Education. Respondent’s degree. (DEGREE)
Family economics. So far as you and your family are concerned, would you say that you
are pretty well satisfied with your present financial situation, more or less satisfied, or not
satisfied at all? (SATFIN)
Family size. Household size and composition. (HOMPOP)
Gender. What is your sex – male or female? (SEX)
Marital status. Are you currently – married, widowed, divorced, separated, or have you
never been married? (MARITAL)
Neighborhood social. How often spend a social evening with someone who lives in your
neighborhood? (SOCOMMUN)
Race. What race do you consider yourself? (RACE)
Subjective well-being. Taken all together, how you would say things are these dayswould you say that you are very happy, pretty happy, or not too happy (HAPPY)
Work status. Last week were you working full time, part time, going to school, keeping
house, or what. (WRKSTAT)

45 
 

APPENDIX B

46 
 

Table B 1
Missing Values for Variables Used in Data Analysis
Variables
Age
Sex
Race
General happiness
Subjective class
Degree
Work status
Marital status
Family economic
Family size
Social contact frequencies
Church attendance

N
331
331
331
328
326
331
331
331
330
331
329
329

Missing
0
0
0
3
5
0
0
0
1
0
2
2

47 
 

APPENDIX C

48 
 

Table C 1
Available Link Functions in the Ordinal Regression Procedure in SPSS
Function

Typical application

Logit

When the dependent ordinal variable has equal category

Complementary log-log

When probability of higher category is high (inverse of N
log-log)

Negative log-log

When probability of lower category is high

Probit

When a dependent variable is normally distributed

Cauchit

Outcome with many extreme values

49 
 

REFERENCES

50 
 

Arthaud-Day, M. L., Rode, J. C., Mooney, C. H., & Near, J. P. (2005). The subjective
well-being construct: a test of its convergent, discriminant, and factorial validity.
Social Indicators Research, 74, 445-476.
Baltes, P. B., & Mayer, K. U. (1999). The Berlin Aging Study: Aging from 70 to100.
New York: Cambridge University Press.
Boyden, P. K., & Aronui, K. K. (2001). Religion and ageing. Colloquium, 33(2), 169180.
Bronfenbrenner, U. (1977). Toward an experimental ecology of human development.
American Psychologist, 32, 513-530.
Bronfenbrenner, U. (1993). The ecology of cognitive development: Research models and
fugitive findings. In R. Wonziak & K. Fischer (Eds.), Development in context:
Acting and thinking in specific environments (pp. 3-44). Hillsdale, NJ: Erlbaum.
Brown, D. (1988). Socio-demographic vs. domain predictors of perceived stress: Racial
differences among American women. Social Indicators Research, 20, 517-532.
Bubolz, M. M. & Sontag, M. S. (1993). Human ecology theory. In P.G. Boss, W.J.
Doherty, R. LaRossa, W.R. Schumm, & S.K. Steinmetz (Eds.), Sourcebook of
family theories and methods: A contextual approach (pp. 419-448). New York:
Plenum Press.
Cantarero, R., Potter, J. J., & Leach, C. K. (2007). Perceptions of quality of life, sense of
community, and life satisfaction among elderly residents in Schuyler and Crete,
Nebraska. Paper presented at the Environmental Design Research Association
38th Annual Conference, Sacramento, CA.
Campbell, A., & Sawer, W. (1976). Life satisfaction in the United States. Social Froces,
54(3), 621-631.
Campbell, A., Converse, P. A., & Rodgers, W. L. (1976). The quality of American life.
New York: Russell Sage Foundation.
Chen, C., & Hughes, J. (2004). Using ordinal regression model to analyze student
satisfaction questionnaires. Association for Institutional Research, 1, 1-13.
Ciarrochi, J.W., & Deneke, E. (2005). Happiness and the varieties of religious
experience: religious support, practices, and spirituality as predictors of wellbeing. Research in the Social Scientific Study of Religion, 15, 209–223.
Clark, A. E., & Oswald, A. (1994). Unhappiness and unemployment. Economic Journal,
104, 648-659.
51 
 

Crooker, K. J., & Near, J. P. (1998). Happiness and satisfaction: Measure of affect and
cognition? Social Indicators Research, 44, 195-224.
Cummins, R. (2002). Subjective well-being from rich and poor, In Glatzer, Wolfgang
(Eds.), Rich and poor: disparities, perceptions, concomintants (pp. 137-156).
Kluwer Academic Publishers Group, Dordrecht, Netherlands.
Diener, Ed. (2000). Subjective well-being. The science of happiness and a proposal for a
national index. American Psychologist, 55 (1), 34-43.
Diener, Ed. (2006). Guidelines for national indicators of subjective well-being and
ill-being. Journal of Happiness Studies, 7, 397-404.
Diener, Ed., & Biswas-Diener, R. (2002). Will money increase subjective well-being? A
literature review and guide to needed research. Social Indicators Research, 57,
119-169.
Diener, Ed., Diener, M., & Diener, C. (1995). Factors predicting the subjective wellbeing of nations. Journal of Personality and Social Psychology, 69(8), 51-64.
Diener, Ed., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The Satisfaction With
Life Scale. Journal of Personality Assessment, 49, 71-75.
Diener, Ed., Gohm, C., Suh, E., & Oishi, S. (2000). Similarity of the relations between
marital status and subjective well-being across cultures. Journal of Cross-Cultural
Psychology, 31, 419-436.
Diener, Ed., & Suh, M. E. (1997). Subjective well-being and age: An international
analysis. Annual review of gerontology and geriatrics, 17, 304-324.
Diener, Ed., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being:
Three decades of progress. Psychological Bulletin, 125, 276-302.
Easterlin, R. A. (2001). Income and Happiness: Towards a Unified Theory. Economic
Journal, 111 (473), 465-484.
Ellison, C. G. (1991). Religious involvement and subjective well-being. Journal
of Health and Social Behavior, 32, 80-99.
Ellison, C., & Kristin L. A. (2001). Religious involvement and domestic violence among
U.S. couples. Journal for the Scientific Study of Religion, 40, 269-287.
 

Fernandez, R., & Kulik, J. (1981). A multilevel model of life satisfaction: Effects of
individual characteristics and neighborhood composition. American Sociological
Review, 12, 840-850.
52 
 

Gana, K., Alaphilippe, D., & Bailly, N. (2004). Positive illusions and mental and physical
health in later life. Aging & Mental Health, 8, 58-64.
George, L. K. (1992). Economic status and subjective well being. In N. Cutler, D. Gregg,
& M. P. Lawton (Eds.), Aging, money and life satisfaction: Aspects of financial
gerontology (pp 69-99). New York: Springer.
George, L. K. (2006). Perceived quality of life. In R. H. Binstock & L. K. George (Eds.),
Handbook of aging and the social science (pp. 320-335). Academic press.
George, L. K. (2010). Still happy after all these years: Research frontiers on subjective
well-being in later life. Journal of Gerontology: Social Science, 65B(3), 331-339.
George, L. K., Okun, M. O., & Landerman, R. (1985). Age as a moderator of the
determinants of life satisfaction. Research on Aging, 7, 209-233.
Griffore, R., Phenice, L. (1988). Causality and the ecology of human development. The
Psychological Record, 38, 515-525.
Griffore, R. & Phenice, L. (2001). The language of human ecology: A general systems
perspective. Dubuque, Iowa: Kendall/Hunt Publishing Co.
Hong, S. M., & Giannakopoulos, E. (1994). The relationship of satisfaction with life to
personality characteristics. The Journal of Psychology, 128(5), 1-8.
Inglehart, R. (2002). Gender, aging, and subjective well-being. International Journal of
Comparative Sociology, 43, 391-408.
Jooste, J. L. (2007). Antecedents of the psychological adjustment of children and
grandparent caregivers in grandparent-headed families. Retrieved from
http://digital.library.unt.edu/data/etd/2007_3/open/meta-dc-5157.tkl
Kashdan, T. B. (2004). The assessment of subjective well-being (issues raised by the
Oxford Happiness Questionnaire). Personality and Individual Differences, 36,
1225-1232.
Kim, J. E., & Moen, P. (2001). Is retirement good or bad for subjective well-being?
American Psychological Society, 10(3), 83-86.
Kirby, S. E., Coleman, P. G., & Daley, D. (2004). Spirituality and well-being in frail and
nonfrail older adults. Journal of Gerontology: Psychological Sciences, 59B, 123–
129.
Krause, N. (1993). Race differences in life satisfaction among aged men and women.
Journals of Gerontology, 48, 235-244.
53 
 

Krause, N. (2003). Religious Meaning and Subjective Well-Being in Late Life.
Journal of Gerontology: Social Sciences, 58(3), 160-170.
Larson, R. (1978). Thirty years of research on the subjective well-being of older
Americans. Journal of Gerontology, 33, 109-125.
Layard, R. (2005). Happiness: Lessons from a New Science. London: Penguin Books.
Loius, V. V. & Zhao, S. (2002). Effects of family structure, family SES, and adulthood
experiences on life satisfaction. Journal of Family Issues, 11, 986-1005.
Lucas, R. E.,Diener, E.,& Suh, E. M. (1996). Discriminant validity of well-being
measures. Journal of Personality and Social Psychology, 71, 616-628.
Lyubomirsky, S., & Lepper, H. S. (1999). A measure of subjective happiness:
preliminary reliability and construct validation. Social Indicators Research, 46,
137-155.
McLaughlin., & Carter, C. (2007). Buying happiness: Family income and adolescent
subjective well-being (Doctoral dissertation). Available from Dissertation Express
database. (UMI No. 3306254)
Mroczek, D. K., & Spiro, A. (2005). Change in life satisfaction during adulthood:
Findings from the Veterans Affairs Normative Aging Study. Journal of
Personality and Social Psychology, 88, 189-202.
Muuss, R. E. (1996). Theories of Adolescence. New York: The McGraw-Hill
Companies, Inc.
Myers, D. G. (1992). The pursuit of happiness. New York: William Morrow.
Myers, D. G. (2008). Religion and human flourishing. In M. Eid, & R. Larsen (Eds.), The
science of subjective well-being (pp. 323–343). New York: Guilford.
National Opinion Research Center. (2006). General social survey archive.
Available at www. sda.berkeley.edu.
Odum, H. T. (1994). Ecological and general systems: An introduction to systems
ecology. Revised edition. Niwot, Colorado: The University Press of Colorado.
Ooms, T. (1996). Where is the family incomprehensive community initiatives for
children and families? Paper presented at the Aspen Round table on
Comprehensive Community Initiatives for Children and Families, Aspen, CO.
Pavot, W. (2008). The assessment of subjective well-being. In M. Eid & R. J. Larson
54 
 

(Eds.), The science of subjective well-being (pp. 124-140). New York: Guilford
Press.
Pearlin, L. I., & Skaff, M. M. (1996). Stress and the life course: A paradigmatic
alliance. The Gerontologist, 36(2), 239-247.
Pranjic, N., Brkovic, A., & Beganlic, A. (2007). Discontent with financial situation,
self-rated health, and well-being of adolescents in Bosnia and Herzagovina:
Cross-sectional study in Tuzla Canton. Croatian Medical Journal, 48(5), 691700.
Pinquart, M., & Sorensen, S. (2000). Influences of socioeconomic status, social network,
and competence on subjective well-being in later life: A meta-analysis.
Psychology and Aging, 15(2), 187-224.
Robine, J. M., Jagger, C., Mathers, C. D., Crimmins, E. M., Suzman, R. M. (2003).
Determining health expectancies. John Wiley and Sons Ltd, Atrium, England.
Rowe J. W., & Kahn R. L. (1998). Successful aging. Pantheon, New York.
Russell, P. D., Burton., Rushing, B., Ritter, C., & Rakocy, A. (1993). Roles, race and
subjective well-being: A longitudinal analysis of elderly men. Social Indicators
Research, 28, 137-156.
Seligman, M. (1988). Boomer blues. Psychology Today, 50-55.
Shrive, F. M., Stuart, H., Quan, H., & Ghali, W. A. (2006). Dealing with missing data in
a multi-question depression scale: a comparison of imputation methods. BMC
Medical Research Methodology, 6, 57.
Smith, J., Fleeson, W., Geiselmann, B., Settersten, R. A. & Kunzmann, U. (1999).
Sources of well-being in very old age. In P.B. Baltes and K. U. Mayer (Eds.), The
Berlin aging study: Aging from 70 to 100. New York: Cambridge University
Press. 450-470.
Smith, J., Maas, I., Mayer, K. U., Helmchen, H., Steinhagen-Thiessen, E., & Baltes, P. B.
(2002). Two-wave longitudinal findings from the Berlin Aging Study:
Introduction to a collection of papers. Journal of Gerontology: Psychological
sciences, 57B, 471-473.
Siahpush, M., Spittal, M., & Singh, G. K. (2008). Happiness and Life Satisfaction
Prospectively Predict Self-Rated Health, Physical Health, and the Presence of
Limiting, Long-Term Health Conditions. American Journal of Health Promotion,
23(1), 18-26.
Streib, G. F., & Beck, R. W. (1980). Older families: A decade review. Journal of
55 
 

Marriage and the Family, 42, 937–956.
Tellegen, A. (1982). Multidimensional personality questionnaire manual. University of
Minnesota Press.
The Pew Research Center. (2006). 2006 generation next study. Retrieved from
http://people-press.org/reports/pdf/300.pdf
U. S. Census Bureau. (2003). Grandparents living with grandchildren: 2000. Retrieved
from http://www.census.gov/prod/2003pubs/c2kbr-31.pdf
Van Hoorn, A. (2007, June). A short introduction to subjective well-being: Its
measurement, correlates and policy uses. Paper presented for
Organization for Economic Co-operation and Development: Conference on
Measuring Progress of Societies. Istanbul.
Veenhoven, R. (1996). Happy life-expectancy: A comprehensive measure of quality-oflife in nations. Social Indicators Research, 39, 1–58.
Veenhoven, R. (2010). How universal is happiness? In E. D. Diener, J. F.
Helliwell & D. Kahneman (Eds.) International Differences in Well-Being (pp
328-350). New York: Oxford University Press.
Warr, P., Butcher, V., & Rovertson, I (2004). Activity and psychological well-being in
older people. Aging Mental Health, 8, 172-183.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect: The PANAS scales. Journal of
Personality and Social Psychology, 54(6), 1063-1070.
Witter, D., Okun, M. A., Stock. W. A., & Haring, M. J. (1984). Education and subjective
well-being: A meta-analysis. Educational Evaluation and Policy Analysis, 6, 165173.
Young, R. D. (2004). Aging in America: With some review of the status of older South
Carolinians. Public Policy and Practice, 2(4), 1540 - 1499.
Yang, Y. (2007). Long and happy living: Trends and patterns of happy life expectancy in
the U.S., 1970-2000. Social Science Research, 37(4), 1235-1252.
Yang, Y. (2008). Social inequalities in happiness in the United States, 1972 to 2004: An
age-period-cohort analysis. American Sociological Review, 73, 204-226.
 

56